Recombination: Holliday Junction Resolution and Crossover Formation Wolf-Dietrich Heyer  Current Biology  Volume 14, Issue 2, Pages R56-R58 (January 2004) DOI: 10.1016/j.cub.2003.12.043

Figure 1 Double-strand-break repair (DSBR) and synthesis-dependent strand annealing (SDSA) models for meiotic recombination. Meiotic recombination is initiated by Spo11-generated double-stranded DNA breaks (DSBs) (step 1). After DSB processing by the Mre11–Rad50–Xrs2 complex and other enzymes (step 2), RPA, Rad52 and Rad55–Rad57 orchestrate formation of the Rad51 nucleoprotein filament capable of homology search and DNA strand invasion (step 3). Rad54 augments Rad51-mediated recombination in D-loop formation and is also thought to allow access of DNA polymerases to the invading 3′-OH by displacing Rad51 from the product heteroduplex DNA (steps 3b, 4a). The D-loop created by single-end invasion (SEI) may enter the DSBR pathway (step 3a) and form a double Holliday junction (dHJ; steps 4a, 5), which can be resolved to crossover (CO; step 5a) and non-crossover (NCO; step 5b) products. Resolution to CO requires a symmetric cleavage of both Holliday junctions in opposite orientations by Holliday junction resolvase. Resolution to NCOs can also be achieved by the resolvase (cleavage of both junctions in the same orientation) and by collapsing the dHJ to a hemi-catenane followed by resolution involving a type I topoisomerase activity. Alternatively, the D-loop enters the SDSA pathway (step 3b). After extension by DNA polymerase, the invading strand retreats to reanneal with the single-stranded DNA tail that did not form a D-loop (step 4b). In its simplest version, SDSA leads only to NCO products (as shown). More complex versions of SDSA involve dHJ formation, which can be resolved by a Holliday junction resolvase to CO and NCO products [3]. Current Biology 2004 14, R56-R58DOI: (10.1016/j.cub.2003.12.043)

Figure 2 A model for crossovers formation without resolution of double Holliday junctions. After single-end invasion, Mus81–Eme1 cleaves the D-loop, which should topologically stabilize single-end invasion. Second-end capture and DNA synthesis produces a nicked Holliday junction subject to a second, possibly independent, cleavage by Mus81–Eme1. The potential Mus81–Eme1/Mms4 cleavage sites are indicated by arrowheads. Processing of the cleavage products to accommodate possible flaps and ligation will always lead to crossovers. Osman et al.[10] propose this alternative crossover model based on the in vitro substrate specificity of Mus81–Eme1. This model was independently [5] derived from the proposed function of XPF–ERCC1 in gene targeting [20]. XPF–ERCC1 is related to Mus81–Eme1/Mms4, the enzymes having a similar subunit structure and sequence homology in their active sites [5]. Current Biology 2004 14, R56-R58DOI: (10.1016/j.cub.2003.12.043)